China’s growing influence in AI is being driven less by headline-grabbing breakthroughs and more by a deliberate push into open-source systems, according to a new warning from a US congressional advisory body.

The report, published by the U.S.-China Economic and Security ⁠Review Commission, points to a shift in how AI power is being accumulated. Rather than competing solely at the frontier of the largest and most expensive models, Chinese firms have focused on distributing adaptable systems that can be widely deployed, modified, and improved over time. “China has opted to go all in on an open-source approach to AI,” the report says.

Major Advantages

That approach, the commission argues, is beginning to create a structural advantage. Open-source models developed by Chinese companies like Alibaba and Moonshot are now widely used across the world. They are cheaper to run, easier to customize, and available to a broader range of users than many closed alternatives from leading US vendors.

The advantage builds over time. As more companies adopt these models, they generate data from real-world applications, ranging from automation to logistics and robotics. “This has resulted in the acceleration of global uptake of Chinese AI and created a feedback loop where widespread adoption drives iteration, then further adoption,” the report says.

Even with restrictions on access to advanced chips, Chinese developers are finding ways to stay competitive by relying on scale of deployment rather than scale of compute.

This strategy is a clear difference between the two countries. American companies, including Anthropic and OpenAI, have invested heavily in cutting-edge models but have become more selective about releasing them openly. Commercial pressures and safety concerns have pushed many of these systems behind APIs, limiting how they can be modified or independently deployed.

Chinese developers, in contrast, release open models designed to run on a range of hardware, making them available to the broader market of engineers and organizations that want control over how AI is used. This strategy has made these systems more accessible to startups and industrial users alike.

As a result, Chinese models now rank among the most downloaded and deployed in open ecosystems, and in some cases have surpassed Western alternatives. Companies outside China are also adopting them, drawn by lower costs and the ability to fine-tune models for specific applications.

The advisory report highlights how China is integrating open AI into physical systems, including manufacturing processes and robotics. These deployments generate continuous streams of operational data, which can be fed back into development. “China’s open AI model strategy and its manufacturing dominance are mutually reinforcing,” the report notes.

Challenges with Oversight

Some Western analysts have raised questions about security risks and potential biases embedded in widely distributed models. Open systems also make oversight more difficult, since they are deployed and modified without centralized control.

Still, adoption continues to grow. For many users, the trade-offs are outweighed by cost and flexibility. The report notes that even companies in Europe and the US are incorporating Chinese open models into their own AI development pipelines.

The broader picture is not one of outright technological supremacy, but of changing leverage. The US still leads in advanced chips and high-end models, supported by enormous investment and vast infrastructure. But China’s emphasis on open distribution is allowing it to expand its presence across the global AI ecosystem.

If that trend continues, influence over AI may depend less on who builds the most advanced systems and more on whose models become the most widely used. Open source, once seen as a secondary track in the AI competition, is now central to how that competition is unfolding.